Why The Very Hard 'Celeste' is Perfectly Fine With You Breaking Its Rules


Celeste is nice, pleasant, and often wants to give you a hug, but at its core, Celeste is a hardcore platformer where you’re meant to die, learn from mistakes, and eventually succeed. It was crafted that way for a reason; it’s what designer Matt Thorson wanted to make. But what if one jump puzzle is too much for your fingers? Do you just give up and move on, defeated? Most of the time, that’s the only course of action. In Celeste, however, there’s Assist Mode, which you can change various game rules—game speed, stamina, the number of air dashes, even your ability to die—and shape the game.

This is a companion discussion topic for the original entry at https://waypoint.vice.com/en_us/article/d3w887/celeste-difficulty-assist-mode


This is a good piece, great to hear a developer talking through their ideas for a11y in a game that’s all about being hard but…

The trend right now seems to be moving in the direction of more choice. If it doesn’t impact the designer-preferred method of playing the game, who’s it harm? […] The question is intent. What does the designer intend for you to experience, to feel, to do? What say does the player have in that process? How much can they change? […] Thorson [] doesn’t believe it ruins authorial intent.

As I’ve said before, this false ableist framing has to go. These absolutely don’t need to be framed as questions. If the question is intent then does this game require me to buy a ~10ms latency gaming display to play it or can I used a normal TV/screen (which displays the game several frames behind this)? Do all versions have the same internal button-press to scan-out latency (note that quite a few games do not maintain this between versions, even with the PS4 and XB1 hardware being similar, there are API/etc differences that mean not every version is identical and this can crop up in major games with huge budgets to get this right)?

If these are not questions that are burning a hole in your mind then there is no singular intent because there is no singular game. There are many various ways in which it can be experienced on various hardware which gives different experiences even before we factor in the variability of the individual playing.

From this perspective: how can a game not have various difficulty modes or assists that can be enabled to give the player the best experience they can have? Surely lacking such options prevents players from engaging with the game design’s intent (either due to a11y issues or hardware/platform issues). Let us flip our ableist language around (in which a11y options have been painted as a potential threat to design intent) and start asking these questions of those who fail to include deep and considered options for assist modes and so on - why do those designers not wish for people to be able to experience their authorial intent via the imperfect medium of complex computer machinery?


I’ve kind of nodded along every time I’ve seen you make this argument before, because it seems reasonable enough on its surface, but I feel compelled to question it today, possibly because I just finished watching Austin’s stream of Getting Over It with Bennett Foddy. Getting Over It is an unusual game in that it’s pretty specifically about being inaccessible and frustrating. However, it’s true that the specific hardware used to play the game changes the game as evidenced by Bennett Foddy himself being surprised by the feel of an expensive mouse when playing the game (it was apparently designed with a cheaper one). Despite the inability to standardize player hardware, the effect of the challenge seems pretty consistent.

I think there are plenty of reasons to be critical of inaccessible games. Even Getting Over It could have made the decision to offer players alternate ways to experience its content. People who have difficulty with mouse-based control schemes may not be able to even experience the intended frustration of the game, when they could if other control options were available. Choosing to impose artificial accessibility barriers to access a game’s content is always going to limit the audience for the game, but I don’t think it can be said that that doesn’t convey intent or that the intent is completely obscured by variability in end-user hardware.

Personally, I think there are plenty of reasons to be critical of the intent that is conveyed by wanting to make a ‘hard’ game. If a player wouldn’t have been motivated to acquire mastery in a particular set of game mechanics without be required to do in order to progress in a game, forcing them to prove mastery isn’t inherently interesting or fun. Then again, having external motivations to engage with mechanics can help players get a lot more out of those systems than they would have otherwise. Pushing players out of their comfort zones while also not pushing beyond where they’re capable or willing to go is a tough design problem. I think it does everyone a disservice to dismiss that problem as intractable or compromised by hardware differences when the issue is about the core player experience.


My reasoning for being so blunt (because being less so hasn’t worked) is simple: this ableist nonsense is the majority of peddled “discourse” around this topic and it’s trash. Why are we happy to just go along with framing things as a11y opposed to artist’s intent? Why do people with different reaction times or a11y needs than the developer get thrown under a bus of “oh well their game is just designed like this, it’s their vision”?

The simplest way of breaking this down is by demanding that vision be purely explained. Any action game says that in response to various stimuli on the screen a window for reactions is required of the player. So tell me the artistic intent. Give it to me in engineering terms as we’re working with machinery here. Exactly how many ms after this does the controller need to get a response before a hard failure?

“Ok, within 220ms of scan-out for this frame, the user needs to have responded.”

“Right, so is that within 210ms of seeing it, which this TV provides, or within 95ms, which is what this other TV forces the user to respond within? Seems like rather a big range of required reaction times you’ve marked down with that spec. Do you mean for the player to be able to (briefly) think before responding or need to react basically automatically because that’s not the same time for every player, even if we force them all to use the same TV.”

If the answer to that is there is no locked exact intent, no specific designed value because it’s up to a host of things involved in the hardware and effectively out of the reach of the developer, then we can be sure that there is no fixed intent we are demolishing by saying that some people have different reaction times to others and so a game needs to be able to account for this. There is no singular vision and that absolutely collapses the argument that offering a range of experiences in any way dilutes the vision (and we can extend this position beyond response windows - the most obvious case where it just doesn’t stand up to scrutiny). There are already a range of experiences offered (something the designer cannot prevent from existing) so to play out this false narrative of “the a11y considerations inherently risk costing the purity of the design” (as comes across in various reporting/podcasting on this topic) is quite frankly hogwash. It’s flat up ableist and not something we should tolerate in progressive communities.

If a game is meant to be frustrating then it should be. But that doesn’t mean that you should ever frame affordances for a11y as inherently compromising that design intent. How do you make a game about that while making it so more people can be frustrated? How can you avoid someone with the best hardware and great reaction times and some rare affinity for what you think of as frustrating from just walking through your game? Now that’s a responsive game design question. Of course it feels different with different hardware, it feels very different with a trackball - this is all part of the wider considerations of the game design (and not thinking about them just shows there is no perfect, untouchable intent that assist modes etc can destroy - the designer has not thought of everything, one thing they need to think about now is inclusivity and marginalised groups they have excluded that they do not need do).

Again, I totally agree that these all are extremely complex design issues. My issue is that this framing of them is a surface-level skimming on the top; that it prevents meaningfully engaging with deeper design questions because it doesn’t understand the issue enough to realise that a11y isn’t the enemy of design intent and that the inherently variable systems we’re all building upon means there is nothing sacred about the precise points any single design picked if it doesn’t offer any flexibility. We’ve all heard the phrase “some people live life on easy mode” to explain systemic privilege; that totally extends to games that lack any affordances to players.

Some people are playing every game (especially the ones without any affordances) on easy and other are playing them on hard; let’s work to ensure the discourse acknowledges that rather than reinforcing notions of design purity requiring many marginalised people be excluded.


I’ve deleted thousands of words in response to this because I can’t seem to stop myself from getting mired in details that I don’t ultimately think are that important. I appreciate your views on this and I think there’s a lot of value in being cautious about the framing.

I think a lot of the problems in The Discourse About Difficulty is that there’s this entanglement between player ability and player preference that is often not explicitly discussed. “Trying to give a marginalized group a play experience similar to that of the designer of the game” is a much different goal than “trying to maximize the number of people who will try some portion of the game by allowing all players to selectively opt out of anything”. The latter goal isn’t bad and it’s not even necessarily incompatible with the first goal, but they’re not equivalent. Celeste’s fine-grained difficulty modifiers allows players who probably could complete the game at its most punishing choose to not do that and it also allows players who would not be able to complete the game under the initial settings to complete the experience as designed.

When Patrick talks about a “designer-preferred method of playing” while also quoting the developer saying that their language was carefully chosen not to imply that requiring assistance makes the experience lesser, there’s certainly room for improvement in framing. In fairness to Patrick though, it seems that some of that conflicting language comes from the developer’s own evolving views from having a “cheat mode” to realizing that that framing wasn’t the best.

The other thing I wanted to say, and this may be getting off topic, is that conflating difficulty with accessibility really minimizes how many things can keep people from playing games. A lot of this discussion about difficulty is rather academic to me because there’s no chance I’m ever going to play the Souls games, not because of difficulty or theme or mechanics, but because they’re 3D games. I’ve read that 10% of the population is affected by non-improving simulation sickness in games and whether or not that’s true, I’m personally extremely wary of anything with a free-moving camera as it’s an almost guaranteed headache. Even watching some Waypoint streams is painful. This isn’t a trait that affects my life in any way except as an inability to enjoy most mainstream games, so I’m certainly not claiming to be someone who is systemically disadvantaged but I have to wonder how many people avoid games because of the ways in which they’re tuned to a specific audience. [It should also be noted that it’s quite easy for me to find games I do enjoy, though they’re not as elaborate or visually expansive as AAA games.]


I’m not really sure why we need to talk about “designer intend”, when a game deliberately has options for players to adjust difficulty/change specific mechanics in the game.

If it’s in there from the start, obviously it’s okay to be used right? And even if it weren’t in there, a lot people are still finding ways to warp whatever you’ve made into something that is more to their liking. Part of what I like about videogames is that on one hand they are designed to provide a specific experience, but that they also let players take ownership over said experience. And who am I, as a designer, to tell people how they should approach my work, especially if said approach makes it less accessible to them?

To give some context: Back in June I released a game that’s also a difficult platformer. Dying over and over again and learning how get through each challenge is kind of an important part of it. Still I went ahead and put a menu with modifiers right at the start, because I thought it’ just easier for everyone, if I give players at least some control over the way they want to enjoy the game.

I can understand that you don’t want cheaters in multiplayer games, because there your enjoyment is also tied to the way people around you behave. But in single player games? Who cares how you get through something? Achievements don’t really count, because those can also be adjusted to accomodate players that use build-in modifiers. For example my game has an Achievement for completing it in Tourism mode. So people who want to 100% it, have to use cheats at least once in their playthrough.

Also generally speaking, if your game completely falls apart when you remove all the difficulty from it, it probably isn’t all that interesting to begin with.


Even if a game was purely about difficulty, that was 100% of the experience, you’d still need to make a game with extensive options because… everyone is different. So even in that case, you’d need to build all these systems by which each player could be met and given the difficult experience that you wished to design. One size doesn’t fit all and, in fact, guarantees that some people get an experience that’s not the hard challenge you wanted while others are potentially given too much to even be possible - not a challenge but just something that’s impossible; which also wasn’t the intended design.

Totally agreed that this weird fetishisation of difficulty and a singular experience (with achievements and removing debug modes from the shipping version etc) isn’t worth it. Let people play solo games however they want. But even if we were to subscribe to that way of thinking, it wouldn’t actually elevate any game that didn’t engage with difficulty to provide each player with their own experience anyway; lacking options doesn’t even get a game closer to that goal.